projects
/
oota-llvm.git
/ history
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
|
commitdiff
|
tree
first ⋅ prev ⋅
next
Add NVCVTFrm (NEON Convert with fractional bits immediate) and modify N2VImm to
[oota-llvm.git]
/
lib
/
Target
/
ARM
/
ARMISelLowering.cpp
2010-03-24
Jim Grosbach
tweak the arm if conversion heuristic
blob
|
commitdiff
|
raw
2010-03-24
Jim Grosbach
try being more permissive for if-conversion on ARM...
blob
|
commitdiff
|
raw
|
diff to current
2010-03-19
Bob Wilson
Revert this change, since it was causing ARM performanc...
blob
|
commitdiff
|
raw
|
diff to current
2010-03-18
Anton Korobeynikov
Get rid of target-specific fp <-> int nodes when still...
blob
|
commitdiff
|
raw
|
diff to current
2010-03-18
Anton Korobeynikov
Get rid of target-specific nodes for fp16 <-> fp32...
blob
|
commitdiff
|
raw
|
diff to current
2010-03-15
Bob Wilson
Translate "cc" clobber in ARM inline assembly to ARM...
blob
|
commitdiff
|
raw
|
diff to current
2010-03-15
Bill Wendling
Now that the default for Darwin platforms is to place...
blob
|
commitdiff
|
raw
|
diff to current
2010-03-14
Anton Korobeynikov
Add substarget feature for FP16
blob
|
commitdiff
|
raw
|
diff to current
2010-03-14
Anton Korobeynikov
Add codegen support for FP16 on ARM
blob
|
commitdiff
|
raw
|
diff to current
2010-03-09
Bill Wendling
The ARM EH experiment worked!
blob
|
commitdiff
|
raw
|
diff to current
2010-03-09
Bill Wendling
This is part of an LLC-beta test used to test <rdar...
blob
|
commitdiff
|
raw
|
diff to current
2010-03-02
Bill Wendling
Remove dead parameter passing.
blob
|
commitdiff
|
raw
|
diff to current
2010-02-24
Bob Wilson
Check for comparisons of +/- zero when optimizing less...
blob
|
commitdiff
|
raw
|
diff to current
2010-02-24
Jim Grosbach
LowerCall() should always do getCopyFromReg() to refere...
blob
|
commitdiff
|
raw
|
diff to current
2010-02-18
Bob Wilson
Use NEON vmin/vmax instructions for floating-point...
blob
|
commitdiff
|
raw
|
diff to current
2010-02-15
David Greene
Remove an assumption of default arguments. This is...
blob
|
commitdiff
|
raw
|
diff to current
2010-02-08
Jim Grosbach
tighten up eh.setjmp sequence a bit.
blob
|
commitdiff
|
raw
|
diff to current
2010-02-02
Evan Cheng
Revert 95130.
blob
|
commitdiff
|
raw
|
diff to current
2010-02-02
Evan Cheng
Pass callsite return type to TargetLowering::LowerCall...
blob
|
commitdiff
|
raw
|
diff to current
2010-01-30
Anton Korobeynikov
Fix a gross typo: ARMv6+ may or may not support unalign...
blob
|
commitdiff
|
raw
|
diff to current
2010-01-27
Evan Cheng
Eliminate target hook IsEligibleForTailCallOptimization.
blob
|
commitdiff
|
raw
|
diff to current
2010-01-19
Bob Wilson
Wrap some comments to 80 columns.
blob
|
commitdiff
|
raw
|
diff to current
2010-01-18
Jim Grosbach
Patch by David Conrad:
blob
|
commitdiff
|
raw
|
diff to current
2010-01-15
Jim Grosbach
Name change for consistency. No functional change.
blob
|
commitdiff
|
raw
|
diff to current
2010-01-15
Jim Grosbach
EmitAtomicCmpSwap() custome inserter needs to delete...
blob
|
commitdiff
|
raw
|
diff to current
2010-01-14
Jakob Stoklund Olesen
ARM "l" constraint for inline asm means R0-R7, also...
blob
|
commitdiff
|
raw
|
diff to current
2010-01-13
Jakob Stoklund Olesen
Fix pasto
blob
|
commitdiff
|
raw
|
diff to current
2009-12-22
Bill Wendling
Add more plumbing. This time in the LowerArguments...
blob
|
commitdiff
|
raw
|
diff to current
2009-12-21
Evan Cheng
Delete the instruction just before the function termina...
blob
|
commitdiff
|
raw
|
diff to current
2009-12-18
Rafael Espindola
Fix libstdc++ build on ARM linux and part of PR5770.
blob
|
commitdiff
|
raw
|
diff to current
2009-12-18
Bob Wilson
Handle ARM inline asm "w" constraints with 64-bit ...
blob
|
commitdiff
|
raw
|
diff to current
2009-12-15
Jim Grosbach
nand atomic requires opposite operand ordering
blob
|
commitdiff
|
raw
|
diff to current
2009-12-14
Jim Grosbach
Add ARMv6 memory and sync barrier instructions
blob
|
commitdiff
|
raw
|
diff to current
2009-12-14
Jim Grosbach
Thumb2 atomic operations
blob
|
commitdiff
|
raw
|
diff to current
2009-12-14
Jim Grosbach
atomic binary operations up to 32-bits wide.
blob
|
commitdiff
|
raw
|
diff to current
2009-12-12
Jim Grosbach
Framework for atomic binary operations. The emitter...
blob
|
commitdiff
|
raw
|
diff to current
2009-12-11
Jim Grosbach
Rough first pass at compare_and_swap atomic builtins...
blob
|
commitdiff
|
raw
|
diff to current
2009-12-10
Jim Grosbach
Add memory barrier intrinsic support for ARM. Moving...
blob
|
commitdiff
|
raw
|
diff to current
2009-12-08
Evan Cheng
- Support inline asm 'w' constraint for 128-bit vector...
blob
|
commitdiff
|
raw
|
diff to current
2009-12-03
Bob Wilson
Recognize canonical forms of vector shuffles where...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-24
Anton Korobeynikov
Materialize global addresses via movt/movw pair, this...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-23
Dan Gohman
Remove ISD::DEBUG_LOC and ISD::DBG_LABEL, which are...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-21
Devang Patel
We are not using DBG_STOPPOINT anymore.
blob
|
commitdiff
|
raw
|
diff to current
2009-11-12
David Greene
Add a bool flag to StackObjects telling whether they...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-12
Evan Cheng
isLegalICmpImmediate should take a signed integer;...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-11
Evan Cheng
Add TargetLowering::isLegalICmpImmediate. It tells...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-09
Jim Grosbach
Use Unified Assembly Syntax for the ARM backend.
blob
|
commitdiff
|
raw
|
diff to current
2009-11-06
Evan Cheng
Remove ARMPCLabelIndex from ARMISelLowering. Use ARMFun...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-03
Bob Wilson
Revert previous change to a comment. The BlockAddresse...
blob
|
commitdiff
|
raw
|
diff to current
2009-11-02
Bob Wilson
Put BlockAddresses into ARM constant pools.
blob
|
commitdiff
|
raw
|
diff to current
2009-11-02
Anton Korobeynikov
Handle splats of undefs properly. This includes the...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-31
Jim Grosbach
Expand 64-bit logical shift right inline
blob
|
commitdiff
|
raw
|
diff to current
2009-10-31
Jim Grosbach
Expand 64-bit arithmetic shift right inline
blob
|
commitdiff
|
raw
|
diff to current
2009-10-31
Jim Grosbach
Expand 64 bit left shift inline rather than using the...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-31
Evan Cheng
It's safe to remat t2LDRpci; Add PseudoSourceValue...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-30
Bob Wilson
Fix a comment.
blob
|
commitdiff
|
raw
|
diff to current
2009-10-30
Rafael Espindola
This fixes functions like
blob
|
commitdiff
|
raw
|
diff to current
2009-10-30
Bob Wilson
Add ARM codegen for indirect branches.
blob
|
commitdiff
|
raw
|
diff to current
2009-10-28
Evan Cheng
Give ARMISD::EH_SJLJ_LONGJMP and EH_SJLJ_SETJMP names.
blob
|
commitdiff
|
raw
|
diff to current
2009-10-28
Evan Cheng
Use fconsts and fconstd to materialize small fp constants.
blob
|
commitdiff
|
raw
|
diff to current
2009-10-21
Bob Wilson
Most of the NEON shuffle instructions do not support...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-21
Evan Cheng
Match more patterns to movt.
blob
|
commitdiff
|
raw
|
diff to current
2009-10-20
Benjamin Kramer
Random #include pruning.
blob
|
commitdiff
|
raw
|
diff to current
2009-10-15
Bob Wilson
Revert svn r80498 and replace it with a different solut...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-13
Bob Wilson
More Neon clean-up: avoid the need for custom-lowering...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-13
Bob Wilson
NEON VLD/VST are now fully implemented. For operations...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-07
Anton Korobeynikov
Add PseudoSourceValues for constpool stuff on ELF ...
blob
|
commitdiff
|
raw
|
diff to current
2009-10-02
Evan Cheng
getFunctionAlignment should return log2 alignment.
blob
|
commitdiff
|
raw
|
diff to current
2009-09-23
Anton Korobeynikov
ARM does not support offset folding (yet). Disable...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-19
Evan Cheng
Fix PR4926. When target hook EmitInstrWithCustomInserte...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-18
Evan Cheng
Enhance EmitInstrWithCustomInserter() so target can...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-16
Bob Wilson
Expand vector floating-point conversions not supported...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-16
Bob Wilson
Expand some more vector operations not supported by...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-16
Bob Wilson
Neon does not support vector divide or remainder. ...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-15
Bob Wilson
Expand all v2f64 arithmetic operations for Neon.
blob
|
commitdiff
|
raw
|
diff to current
2009-09-09
Bob Wilson
Fix pr4939: Change FPCCToARMCC to translate SETOLE...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-08
Anton Korobeynikov
Unbreak getOnesVector() / getZeroVector() to use valid...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-03
Evan Cheng
Reference to hidden symbols do not have to go through...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-02
Sandeep Patel
Retype from unsigned to CallingConv::ID accordingly...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-01
Bob Wilson
Add support for generating code for vst{234}lane intrin...
blob
|
commitdiff
|
raw
|
diff to current
2009-09-01
Bob Wilson
Generate code for vld{234}_lane intrinsics.
blob
|
commitdiff
|
raw
|
diff to current
2009-09-01
Jim Grosbach
Clean up LSDA name generation and use for SJLJ exceptio...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-30
Anton Korobeynikov
EXTRACT_VECTOR_ELEMENT can have result type different...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-29
Anton Korobeynikov
Do not assert on too wide splats we don't support.
blob
|
commitdiff
|
raw
|
diff to current
2009-08-28
Evan Cheng
Let Darwin linker auto-synthesize stubs and lazy-pointe...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-27
Anton Korobeynikov
Hopefully the final missing part :(
blob
|
commitdiff
|
raw
|
diff to current
2009-08-27
Anton Korobeynikov
Transform float scalar_to_vector into subreg accesses.
blob
|
commitdiff
|
raw
|
diff to current
2009-08-26
Bob Wilson
Remove unneeded ARM-specific DAG nodes for VLD* and...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-26
Anton Korobeynikov
Expand scalar_to_vector - we don't have any isel logic...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-22
Eli Friedman
Make x86 test actually test x86 code generation. Fix the
blob
|
commitdiff
|
raw
|
diff to current
2009-08-21
Bob Wilson
Match VTRN, VZIP, and VUZP shuffles. Restore the tests...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-21
Anton Korobeynikov
Fix some typos and use type-based isel for VZIP/VUZP...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-21
Anton Korobeynikov
Add lowering of ARM 4-element shuffles to multiple...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-21
Anton Korobeynikov
Add nodes & dummy matchers for some v{zip,uzp,trn}...
blob
|
commitdiff
|
raw
|
diff to current
2009-08-21
Anton Korobeynikov
Expand EXTRACT_SUBVECTOR
blob
|
commitdiff
|
raw
|
diff to current
2009-08-21
Anton Korobeynikov
Provide vext.{16,32}
blob
|
commitdiff
|
raw
|
diff to current
2009-08-21
Anton Korobeynikov
Use masks not nodes for vector shuffle predicates....
blob
|
commitdiff
|
raw
|
diff to current
2009-08-19
Bob Wilson
Add support for Neon VEXT (vector extract) shuffles.
blob
|
commitdiff
|
raw
|
diff to current
2009-08-15
Bill Wendling
Reapply r79127. It was fixed by d0k.
blob
|
commitdiff
|
raw
|
diff to current
2009-08-15
Bill Wendling
Revert r79127. It was causing compilation errors.
blob
|
commitdiff
|
raw
|
diff to current
next